Data Center Energy Surge Exposes Natural Gas Power Plants to Sudden Price Volatility

The electricity demand from data centers is not a minor uptick; it represents a fundamental shift in the U.S. energy landscape. For over a decade, national power consumption was essentially flat. That pattern has broken, with demand now growing at an annual rate of 1.7% since 2020. This marks a clear divergence from the past and sets the stage for a supply-constrained shock.

The scale of the new demand is staggering. Projections indicate that domestic data center energy use will double or triple by 2028. This isn't just growth; it's a structural acceleration. The U.S. Department of Energy frames this as a return to the rising demand trends last seen before the early 2000s, driven by a growing economy and consumer adoption. Now, that cycle is restarting, powered by artificial intelligence and digital infrastructure. The implications are broad: total energy demand could grow by 15-20% in the next decade.

This surge is already visible in the data. In 2023, data centers consumed about 4.4% of total U.S. electricity. By 2028, that share is expected to climb to between 6.7% and 12%, with total usage potentially reaching 325 to 580 TWh. The EIA's latest forecast underscores the pressure, projecting nationwide electricity load will increase by 1.9% in 2026 and 2.5% in 2027. This isn't a cyclical blip. It's a sustained, supply-demand imbalance that is tightening power supplies and setting the stage for higher costs and greater reliance on fossil fuels in the near term.

The Supply Response and Market Pressures

The power generation sector is scrambling to respond to this new demand, but the path is narrow and costly. The primary mechanism is a rapid shift toward natural gas, which is seen as the most flexible fuel to cover the additional load. Under a high-demand scenario that assumes a faster data center build-out, the EIA projects natural gas-fired electricity output will rise by 7.3% between 2025 and 2027. That's a stark contrast to the 1.7% increase forecast under the agency's baseline scenario. This accelerated reliance on gas is the direct supply response to the demand shock, but it comes with immediate price consequences.

The mechanism is straightforward: increased demand for gas to run power plants pushes up fuel costs, which are then passed through to wholesale electricity prices. The EIA forecasts these price increases will be most severe in the regions with the fastest load growth. The Texas grid operator ERCOT is projected to see the steepest climb, with its 2027 prices potentially averaging $37/MWh higher than its February projections-a rise of 79%. This extreme volatility is tied to ERCOT's limited ability to import power from neighboring regions during peak demand, creating a local bottleneck. Other regions could face more moderate but still significant hikes, with the EIA noting wholesale power prices could climb by around $1–$3/MWh in some areas.

This regional strain highlights a critical vulnerability. The fastest-growing load is concentrated in specific grid areas, particularly ERCOT and PJM. Their infrastructure is being pushed to the limit, and the supply response of adding more gas capacity may not keep pace with demand spikes. The result is a market under acute pressure, where the cost of electricity is set to rise faster than the broader economy can easily absorb. For now, the fuel mix is shifting to meet the immediate need, but the price signal is clear: the commodity balance is tightening, and the cost of maintaining that balance is rising.

Policy and Investment Catalysts

The market pressures are now spurring a wave of policy and investment action, aiming to accelerate supply or shift capital expenditure away from the grid. These initiatives are critical for bridging the gap, but their impact will depend on execution speed and scale.

The Department of Energy is taking a direct, infrastructure-led approach. It has released a Request for Information (RFI) to explore co-locating data centers with new energy generation on federal lands. The plan targets 16 specific sites with existing energy infrastructure, aiming to fast-track permitting. The goal is ambitious: to have AI infrastructure commencing operation by the end of 2027. This strategy directly addresses the supply-demand mismatch by creating dedicated zones where power and data can be built in tandem, potentially reducing transmission bottlenecks. The focus on public-private partnerships and leveraging national labs suggests a push for rapid deployment of innovative technologies, including nuclear.

At the same time, a major policy shift is being directed at the demand side. President Trump announced a "ratepayer protection pledge" during his State of the Union, directing tech companies to pay for their own power generation for new data centers. This "bring your own generation" model aims to shift the capital expenditure burden from utilities and ratepayers to the companies consuming the power. The pledge is expected to be formalized in a White House meeting on March 4, 2026, with major tech firms in attendance. While the specifics are still emerging, the directive could accelerate private investment in on-site generation and storage, providing a more distributed supply response.

Complementing these federal moves, state governments are acting on a massive scale. From 2021 to 2025, lawmakers passed more than 400 DER-related legislative actions. These policies focus on distributed energy resources like rooftop solar and battery storage, using tools such as investment support and net metering to incentivize adoption. By streamlining interconnection and improving grid planning, states are creating a framework to integrate these resources as a reliable source of capacity. This decentralized approach can help defer costly utility-scale transmission and generation projects, providing a faster, more flexible supply response to local demand spikes.

The combined effect of these catalysts is a multi-pronged effort to rewire the supply chain. The DOE's land-based projects target large-scale, centralized solutions. The ratepayer pledge pushes tech giants to fund their own distributed generation. And state policies create a broad ecosystem for distributed resources. Together, they aim to accelerate the supply response, but the timeline for meaningful impact remains tight, with the DOE's target of end-of-2027 operation representing a critical early benchmark.

Catalysts and Risks to the Balance

The current trajectory of energy demand and supply is now entering a phase where visibility and execution will be the critical tests. The coming weeks will provide a key signal on the true scale of the problem, while the path forward faces two major risks that could validate or challenge the market's current pricing.

The most immediate catalyst is the U.S. Department of Energy's pilot survey, which begins this week. The Energy Information Administration is launching a series of pilot surveys on Wednesday to gauge the energy use of data centers, starting in Virginia, Washington State, and Texas. This is a direct response to the lack of reliable data, which has left analysts and policymakers operating in the dark. The purpose is clear: to move from projections to verified facts. The survey will ask about backup power supplies and fuel types, aiming to build a "patchwork quilt of lots of different things" that can inform future policy and investment. For the market, this is a validation signal. If the survey confirms the high-end demand scenarios, it will reinforce the need for the supply responses already being planned. If it finds the demand is lower, it could ease near-term price pressures. Either way, it provides the missing data point that has been absent from the analysis.

The primary risk to the supply-demand balance is that new generation capacity simply cannot keep pace. The EIA's high-demand scenario assumes a 7.3% increase in natural gas-fired output between 2025 and 2027 to cover the load. But this assumes a smooth build-out of gas plants, nuclear reactors, and renewables. If permitting delays, material shortages, or financing issues slow this deployment, the price signal will intensify. The current forecast already shows extreme volatility, with ERCOT prices potentially rising $37/MWh higher by 2027. A supply shortfall would exacerbate this, leading to more frequent and severe price spikes, and raising reliability concerns as grids strain. The trajectory depends on this supply response materializing on schedule.

A parallel risk is public opposition to data center siting, which could slow permitting and deployment. A recent Pew survey found that while Americans see economic benefits, attitudes are more negative than positive when it comes to data centers' impact on the environment and home energy costs. This public sentiment, combined with concerns about local strain on power and water resources, creates a permitting bottleneck. If communities successfully block or delay projects, it could force a reliance on more expensive and less efficient solutions, or simply fail to meet the demand. This risk is a direct challenge to the current trajectory, as it could decelerate the very build-out that is driving the energy demand forecast in the first place.

The bottom line is that the coming months will test the assumptions underpinning the energy market. The DOE's survey provides a crucial data point to validate the demand scale. Meanwhile, the execution of the supply response and the ability to navigate public opposition will determine whether the market's price volatility is a temporary shock or the new normal.